You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@qpid.apache.org by Jin <ji...@titletrading.com> on 2010/06/02 15:24:46 UTC
Performance : Qpid vs 29west UMDS , just for referenced , not for
benchmark.
Hi All:
I did a perf test on May/19/2010 from Qpid trunk.
hardware : intel i920 4G RAM
OS: linux 2.6.32.
GCC: 4.5.0
on same system, the 29west UMDS's performance:
This test is just for internal reference. Please don't put it into
public access.
Jinius.
Re: Performance : Qpid vs 29west UMDS , just for referenced , not
for benchmark.
Posted by Carl Trieloff <cc...@redhat.com>.
On 06/02/2010 09:24 AM, Jin wrote:
>
> Hi All:
>
> I did a perf test on May/19/2010 from Qpid trunk.
>
> hardware : intel i920 4G RAM
>
> OS: linux 2.6.32.
>
> GCC: 4.5.0
>
>
>
> on same system, the 29west UMDS's performance:
>
>
> This test is just for internal reference. Please don't put it into
> public access.
>
> Jinius.
Another thing to play with is that umdssend and perftest with default
settings are not doing the same thing.
Carl.
Re: Performance : Qpid vs 29west UMDS , just for referenced , not for
benchmark.
Posted by Martin Ritchie <ri...@apache.org>.
On 2 June 2010 15:02, Rajith Attapattu <ra...@gmail.com> wrote:
> On Wed, Jun 2, 2010 at 9:51 AM, Martin Ritchie <ri...@apache.org> wrote:
>> On 2 June 2010 14:42, Emmanuel Bourg <eb...@apache.org> wrote:
>>> Just out of curiosity, is it possible to run the same test against the Java
>>> broker to compare the performance?
>>
>> If you do have time it would be good if you could run it on the 0-91
>> protocol as well as the 0-10. 0-10 support is new and so not as well
>> tested.
>>
>> You can do this by excluding 0-10 support at startup.
>>
>> qpid-server -p 5672 --exclude-0-10 5672
>>
>
> The c++ perftest will only work with 0-10 protocol.
> So you really can't run that against the 0-91 version.
Ah sorry didn't notice it was a cpp test.
>> Cheers
>>
>> Martin
>>
>>
>>> Emmanuel Bourg
>>>
>>>
>>>
>>> Le 02/06/2010 15:24, Jin a écrit :
>>>>
>>>> Hi All:
>>>>
>>>> I did a perf test on May/19/2010 from Qpid trunk.
>>>>
>>>> hardware : intel i920 4G RAM
>>>>
>>>> OS: linux 2.6.32.
>>>>
>>>> GCC: 4.5.0
>>>>
>>>>
>>>>
>>>> on same system, the 29west UMDS's performance:
>>>>
>>>>
>>>> This test is just for internal reference. Please don't put it into
>>>> public access.
>>>>
>>>> Jinius.
>>>
>>>
>>>
>>
>>
>>
>> --
>> Martin Ritchie
>>
>> ---------------------------------------------------------------------
>> Apache Qpid - AMQP Messaging Implementation
>> Project: http://qpid.apache.org
>> Use/Interact: mailto:dev-subscribe@qpid.apache.org
>>
>>
>
>
>
> --
> Regards,
>
> Rajith Attapattu
> Red Hat
> http://rajith.2rlabs.com/
>
> ---------------------------------------------------------------------
> Apache Qpid - AMQP Messaging Implementation
> Project: http://qpid.apache.org
> Use/Interact: mailto:dev-subscribe@qpid.apache.org
>
>
--
Martin Ritchie
---------------------------------------------------------------------
Apache Qpid - AMQP Messaging Implementation
Project: http://qpid.apache.org
Use/Interact: mailto:dev-subscribe@qpid.apache.org
Re: Performance : Qpid vs 29west UMDS , just for referenced , not for
benchmark.
Posted by Rajith Attapattu <ra...@gmail.com>.
On Wed, Jun 2, 2010 at 9:51 AM, Martin Ritchie <ri...@apache.org> wrote:
> On 2 June 2010 14:42, Emmanuel Bourg <eb...@apache.org> wrote:
>> Just out of curiosity, is it possible to run the same test against the Java
>> broker to compare the performance?
>
> If you do have time it would be good if you could run it on the 0-91
> protocol as well as the 0-10. 0-10 support is new and so not as well
> tested.
>
> You can do this by excluding 0-10 support at startup.
>
> qpid-server -p 5672 --exclude-0-10 5672
>
The c++ perftest will only work with 0-10 protocol.
So you really can't run that against the 0-91 version.
> Cheers
>
> Martin
>
>
>> Emmanuel Bourg
>>
>>
>>
>> Le 02/06/2010 15:24, Jin a écrit :
>>>
>>> Hi All:
>>>
>>> I did a perf test on May/19/2010 from Qpid trunk.
>>>
>>> hardware : intel i920 4G RAM
>>>
>>> OS: linux 2.6.32.
>>>
>>> GCC: 4.5.0
>>>
>>>
>>>
>>> on same system, the 29west UMDS's performance:
>>>
>>>
>>> This test is just for internal reference. Please don't put it into
>>> public access.
>>>
>>> Jinius.
>>
>>
>>
>
>
>
> --
> Martin Ritchie
>
> ---------------------------------------------------------------------
> Apache Qpid - AMQP Messaging Implementation
> Project: http://qpid.apache.org
> Use/Interact: mailto:dev-subscribe@qpid.apache.org
>
>
--
Regards,
Rajith Attapattu
Red Hat
http://rajith.2rlabs.com/
---------------------------------------------------------------------
Apache Qpid - AMQP Messaging Implementation
Project: http://qpid.apache.org
Use/Interact: mailto:dev-subscribe@qpid.apache.org
Re: Performance : Qpid vs 29west UMDS , just for referenced , not for
benchmark.
Posted by Martin Ritchie <ri...@apache.org>.
On 2 June 2010 14:42, Emmanuel Bourg <eb...@apache.org> wrote:
> Just out of curiosity, is it possible to run the same test against the Java
> broker to compare the performance?
If you do have time it would be good if you could run it on the 0-91
protocol as well as the 0-10. 0-10 support is new and so not as well
tested.
You can do this by excluding 0-10 support at startup.
qpid-server -p 5672 --exclude-0-10 5672
Cheers
Martin
> Emmanuel Bourg
>
>
>
> Le 02/06/2010 15:24, Jin a écrit :
>>
>> Hi All:
>>
>> I did a perf test on May/19/2010 from Qpid trunk.
>>
>> hardware : intel i920 4G RAM
>>
>> OS: linux 2.6.32.
>>
>> GCC: 4.5.0
>>
>>
>>
>> on same system, the 29west UMDS's performance:
>>
>>
>> This test is just for internal reference. Please don't put it into
>> public access.
>>
>> Jinius.
>
>
>
--
Martin Ritchie
---------------------------------------------------------------------
Apache Qpid - AMQP Messaging Implementation
Project: http://qpid.apache.org
Use/Interact: mailto:dev-subscribe@qpid.apache.org
Re: Performance : Qpid vs 29west UMDS , just for referenced , not
for benchmark.
Posted by Emmanuel Bourg <eb...@apache.org>.
Just out of curiosity, is it possible to run the same test against the
Java broker to compare the performance?
Emmanuel Bourg
Le 02/06/2010 15:24, Jin a écrit :
> Hi All:
>
> I did a perf test on May/19/2010 from Qpid trunk.
>
> hardware : intel i920 4G RAM
>
> OS: linux 2.6.32.
>
> GCC: 4.5.0
>
>
>
> on same system, the 29west UMDS's performance:
>
>
> This test is just for internal reference. Please don't put it into
> public access.
>
> Jinius.
Re: Performance : Qpid vs 29west UMDS , just for referenced , not for
benchmark.
Posted by Rajika Kumarasiri <ra...@wso2.com>.
Jin,
I think it'd be a good idea to share your performance testing framework(so
that qpid can perform a similar test), on the other hand I think that the
frame work is not that complex to share, which just send number of messages
to particular queue/topic and log the stats( at least that what I
understood from the out puts).
Rajika
On Wed, Jun 2, 2010 at 6:54 PM, Jin <ji...@titletrading.com> wrote:
> Hi All:
>
> I did a perf test on May/19/2010 from Qpid trunk.
>
> hardware : intel i920 4G RAM
>
> OS: linux 2.6.32.
>
> GCC: 4.5.0
>
>
> on same system, the 29west UMDS's performance:
>
>
> This test is just for internal reference. Please don't put it into public
> access.
>
> Jinius.
>
--
http://rajikak.blogspot.com/
Re: Performance : Qpid vs 29west UMDS , just for referenced , not for
benchmark.
Posted by Martin Ritchie <ri...@apache.org>.
On 2 June 2010 14:24, Jin <ji...@titletrading.com> wrote:
> Hi All:
> This test is just for internal reference. Please don't put it into public
> access.
>
You do realise this is a publicly accessible list? Even if the numbers are
not easily reachable.
> Jinius.
>
--
Martin Ritchie